Quick and Dirty Guide to C

The single best book on C is The C Programming Language by Kernighan and Richie. FUNCTIONS:
A function is a pointer to some code, parameterized by formal parameters, that
CODE: may be executed by providing actual parameters. Functions must be declared before
Code for execution goes into files with “.c” suffix. they are used, but code may be provided later. A sqrt function for positive n
Shared decl’s (included using #include “mylib.h”) in “header” files, end in “.h” might be declared as:
double sqrt(double n) {
COMMENTS: double guess;
Characters to the right of // are not interpreted; they’re a comment. for (guess = n/2.0; abs(n-guess*guess)>0.001; guess = (n/guess+guess)/2);
Text between /* and */ (possibly across lines) is commented out. return guess;
}
DATA TYPES: This function has type double (s*sqrt)(double).
Name! ! Size! Description printf(“%g\n”, sqrt(7.0)); //calls sqrt; actuals are always passed by value
char! ! 1 byte! an ASCII value: e.g. ‘a’ (see: man ascii) Functions parameters are always passed by value. Functions must return a value.
int/long! 4 bytes !a signed integer: e.g. 97 or hex 0x61, oct 0x141 The return value need not be used. Function names with parameters returns the
long long ! 8 bytes! a longer multi-byte signed integer function pointer. Thus, an alias for sqrt may be declared:
float! ! 4 bytes! a floating-point (possibly fractional) value double (*root)(double) = sqrt;
double!! 8 bytes !a double length float printf(“%g\n”, root(7.0));
char, int, and double are most frequently and easily used in small programs Procedures or valueless functions return ‘void’.
sizeof(double) computes the size of a double in addressable units (bytes) There must always be a main function that returns an int.
Zero values represent logical false, nonzero values are logical true. int main(int argc, char **argv) OR int main(int argc, char *argv[])
Math library (#include <math.h>, compile with -lm) prefers double. Program arguments may be accessed as strings through main’s array argv with argc
elements. First is the program name. Function declarations are never nested.
CASTING:
Preceding a primitive expression with an alternate parenthesized type converts or OPERATIONS:
“casts” value to a new value equivalent in new type: +, -, *, /, %! Arithmetic ops. /truncates on integers, % is remainder.
int a - (int) 3.131; ! //assigns a=3 without complaint ++i --i!! Add or subtract 1 from i, assign result to i, return new val
Preceding any other expression with a cast forces new type for unchanged value. i++ i--!! Remember i, inc or decrement i, return remembered value
double b = 3.131; && || !!! Logical ops. Right side of && and || unless necessary
int a = *(int*)&b; //interprets the double b as an integer (not necessarily 3) & | ^ ~!! Bit logical ops: and, or, xor, complement.
>> <<! ! Shift right and left: int n=10; n <<2 computes 40.
STRUCTS and ARRAYS and POINTERS and ADDRESS COMPUTATION: = ! ! Assignment is an operator. Result is value assigned.
Structs collect several fields into a single logical type: += -= *= etc ! Perform binary op on left and right, assign result to left
struct { int n; double root;} s; //s has two fields, n and root == != < > <= >= Comparison operators (useful only on primitive types)
s.root = sqrt((s.n=7)); //ref fields (N.B. double parens=>assign OK!) ?:! ! If-like expression: (x%2==0)?”even”:”odd”
Arrays indicated by right associative brackets ([]) in the type declaration ,! ! computing value is last: a, = b,c,d; exec’s b,c,d then a=d
int a[10]; //a is a 10int array. a[0] is the first element. a[9] is the last
char b[]; //in a function header, b is an array of chars with unknown length STATEMENTS:
int c[2][3]; //c is an array of 2 arrays of three ints. a[1][0] follows a[0][2] Angle brackets identify syntactic elements and don’t appear in real statements
Array variables (e.g. a,b,c above) cannot be made to point to other arrays <expression> ;! ! //semicolon indicates end of a simple statement
Strings are represented as character arrays terminated by ASCII zero. break;!! ! //quits the tightest loop or switch immediately
Pointers are indicated by left associative asterisk (*) in the type declarations: continue;! ! //jumps to next loop test, skipping rest of loop body
int a*a; // a is a pointer to an integer return x;! ! //quits this function, returns x as value
char *b; // b is a pointer to a character { <statements> } ! //curly-brace groups statements into 1 compound (no ;)
int *c[2]; // c is an array of two pointers to ints (same as int *(c[2]); if (<condition>) <stmt>!//stmt executed if cond true (nonzero)
int (*d)[2]; // d is a pointer to an array of 2 integers if (<condition>) <stmt> else <stmt> // two-way condition
Pointers are simply addresses. Pointer variables may be assigned. while (<condition>) <stmt> //repeatedly execute stmt only if condition true
Adding 1 computes pointer to the next value by adding sizeof(X) for type X do <stmt> while (<condition>); //note the semicolon, executes at least once
General int adds to pointer (even 0 or negative values) behave in the same way for (<init>; <condition>; <step>) <statement>
Addresses may be computed with the ampersand (&) operator.
An array without an index or a struct without field computes its address: switch (<expression>) { ! //traditional “case statement”
int a[10], b[20]; // two arrays case <value>: <statement>! // this statement exec’d if val==expr
int *p = a; // p points to first int of array a ! ! break;! ! // quit this when value == expression
p = b; // p now points to the first int of array b case <value2>: <statement2>! //executed if value2 = expression
An array or pointer with an index n in square brackets returns the nth value: case <value3>: <statement3> //executed if value3 = expression
int a[10]; // an array ! ! break;! ! // quit
int *p; default: <statement4>!! // if matches no other value; may be first
int i = a[0]; // i is the first element of a ! ! break;! ! // optional (but encouraged) quit
i = *a; // pointer dereference }
p = a; // same as p = &a[0]
p++; // same as p = p+1; same as p=&a[1]; same as p = a+1 KEY WORDS
Bounds are not checked; your responsibility not to run off. Don’t assume. unsigned! before primitive type suggests unsigned operations
An arrow (-> no spaces!) dereferences a pointer to a field: extern!! in global declaration => symbol is for external use
struct { int n; double root; } s[1]; //s is pointer to struct or array of 1 static!! in global declaration => symbol is local to this file
s->root = sqrt)s->n = 7); //s->root same as (*s).root or s[0].root ! ! in local decl’n => don’t place on stack; keep value betw’n calls
printf(“%g\n”, s->root); typedef! before declaration defines a new type name, not a new variable
Quick and Dirty Guide to C Content borrowed and updated (with permission)
from Duane A. Bailey’s guidelines from 2007.

I/O (#include <stdio.h>)

Default input comes from “stdin”; output goes to “stdout”; errors to “stderr”.
Standard input and output routines are declared in stdio.h: #include <stdio.h> A GOOD FIRST PROGRAM:
Function! Description #include <stdio.h>
fopen(name, “r”) opens file name for read, returns FILE *f; “w” allows write #include <stdlib.h>
fclose(f)! closes file f int main(int argc, char** argv){
getchar()! read 1 char from stdin or pushback; is EOF (int -1) if none printf(“Hello, world.\n”);
ungetch(c)! pushback char c into stdin for re-reading; don’t change c return 0;
putchar(c)! write 1 char, c, to stdout }
fgetc(f)! same as getchar(), but reads from file f
ungetc(c,f)! same as ungetchar() but onto file f A WORD COUNT (WC)
fputc(c,f)! same as putchar(c), but onto file f #include <stdio.h>
fgets(s,n, f)! read string of n-1 chars to a s from f or til eof or \n #include <stdlib.h>
fputs(s,f)! writes string s to f: e.g. fputs(“Hello world\n”, stdout); int main(int argc, char **argv){
scanf(p,...) ! reads ... args using format p (below); put &w/non-pointers int charCount=0, wordCount=0, lineCount=0;
printf(p, ...) write ... args using format p (below); pass args as is int doChar=0, doWord=0, doLine=0, inWord = 0;
fprintf(f,p,...) same, but print to file f int c;
fscanf(f,p,...) same, but read from file f char *fileName = 0;
sscanf(s,p,...) same, but read from string s FILE *f = stdin;
sprintf(s,p,...) same, as printf, but to string s while (argv++, --argc) {
feof(f)! return true iff at end of file f ! if (!strcmp(*argv,”-c”)) doChar=1;
Formats use format characters preceded by escape %; other chars written as is> ! else if (!strcmp(*argv,”-w”)) doWord=1;
char! meaning! ! ! ! char! meaning ! else if (!strcmp(*argv,”-l”)) doLine=1;
%c! character! ! ! \n! newline (control-j) ! else if (!(f = fopen((fileName = *argv), “r”))){
%d! decimal integer! ! ! \t ! tab (control-i) ! printf(“Usage: wc [-l] [-w] [-c]\n”); return 1;
%s! string! ! ! ! \\! slash ! }
%g ! general floating point! ! %%! perent }
if (!(doChar || doWord || doLine)) doChar = doWord = doLine = 1;
MEMORY (%include <stdlib.h>) while (EOF != (c= fgetc(f))){
malloc(n)! alloc n bytes of memory; for type T: p = (T*)malloc(sizeof(t)); ! charCount++;
free(p)! free memory pointed at p; must have been alloc’d; don’t re-free ! if (c == ‘\n’) lineCount++;
calloc(n,s)! alloc n-array size s & clear; typ: a = (T*)calloc(n, sizeof(T)); ! if (!iswpace(c)) {
! if (!inWord) { inWord = 1; wordcount++; }
MATH (#include <math.h> and link -lm; sometimes documented in man math) ! } else { inWord = 0; }
All functions take and return double unless otherwise noted: }
sin(a), cos(a), tan(a)! sine, cosine, tangent of double (in radians) if (doLine) printf(“%8d”, lineCount);
asine(y),acos(x),atan(r) principle inverse of above if (doWord) printf(“%8d”, wordCount);
atan2(y,x)! ! principal inverse of tan(y/x) in same quadrant as (x,y) if (doChar) printf(“%8d”, charCount);
sqrt(x)! ! root of x if (fileName) printf(“ %s”, fileName);
log(x)!! ! natural logarithm of x; others: log2(x) and log10(x) printf(“\n”);
exp(p)!! ! e to the power of p; others: exp2(x) and exp10(x) }
pow(x,y)! ! x to the power of y; like (expy*log(x))
ceil(x)! ! smallest integer (returned as double) no less than x
floor(x)! ! largest integer (returned as double) no greater than y ADD YOUR NOTES HERE:
#include <stdlib.h> for these math functions
abs(x)!! ! absolute value of x
random()! ! returns a random long
srandom(seed)! ! seeds the random generator with a new random seed

STRINGS (#include <string.h>)

strlen(s)! return length of string; number of characters before ASCII 0
strcpy(d,s)! copy string s to d and return d; N.B. parameter order like =
strncpy(d,s,n)! copy at most n characters of s to d and terminate; returns d
stpcpy(d,s)! like strcpy, but returns pointer to ASCII 0 terminarot in d
strcmp(s,t)! compare strings s and t and return first difference; 0=> equal
strncmp(s,t,n)! stop after at most n characters; needn’t be null terminated
memcpy(d,s,n) copy exactly n bytes from s to d; may fail if s overlaps d
memmove(d,s,n) (slow) copy n bytes from s to d; won’t fail if s overlaps d

COMPILING:
gcc prog.c! # compiles prog.c into a.out run result with ./a.out
gcc -o prog prog.c # compiles prog.c into prog; run result with ./prog
gcc -g -o prog prog.c # as above, but allows for debugging